st: New package -scdensity- available from SSC

[Joerg Luedicke <joerg.luedicke@gmail.com>]

Hi everybody,

Thanks to Kit Baum, a new package -scdensity- is available from SSC.
Type -ssc install scdensity- to download and install it, then type
-help scdensity- for further information. -scdensity- requires Stata
version 9.2 or higher and Ben Jann's -moremata- package, also
available from SSC (type -ssc install moremata- to install it).

-scdensity- is an implementation of the self-consistent density
estimator as described in Bernacchia & Pigolotti (2011). The
self-consistent method is a new method for non-parametric density
estimation. Unlike other popular density estimators, the
self-consistent methods estimates the density of a continuous
probability distribution without any prior assumptions or parameter
fixations (like smoothing parameters in classical kernel estimates,
for example). The basic idea of the self-consistent estimator is to
iteratively derive an optimally shaped kernel, instead of searching
for an optimal bandwidth, given a kernel function (like it is done in
classical kernel density estimation).

Simulation results show that the self-consistent method is performing
very well, especially for larger sample sizes (i.e., N>1000). I will
present this program next week (July 26-27, 2012) at the Stata
Conference in San Diego (including results from the Monte Carlo
simulations).  After the conference, presentations will be uploaded to
stata.com and repec.org, so if you are curious about the accuracy of
the new estimator, have a look at the presentation then.

It is likely that -scdensity- will be extended in the near future. The
next step will be to allow for 2-dimensional density estimation, so
watch out for updates. As of now, -scdensity- performs univariate
density estimation only.

If you have any questions about the program, feel free to contact me
(please use the email address that is given in the help file).

All the best,

Joerg


References:

Bernacchia, A. & Pigolotti, S. (2011). Self-Consistent Method For
Density Estimation.
Journal of the Royal Statistical Society B, 73(3), 407-422.
*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/